Washburn Lab

The goal of our research is to develop label free quantitative proteomic tools with a particular focus on applying these tools to the analysis of transcriptional regulatory complexes. On the technology side, we are focused on the further development of spectral counting as a quantitative proteomic tool. On the biological side, we are particularly interested in the discovery of novel protein-protein interactions with well characterized transcriptional regulatory complexes, the proteomic based analysis of the dynamics of transcriptional regulatory complexes, and characterization of protein interaction networks. We are a highly interactive and interdisciplinary group, and we collaborate extensively with many researchers at the Stowers Institute.

 

Multidimensional Protein Identification Technology (MudPIT)

We utilize a method named Multidimensional Protein Identification Technology (MudPIT) to analyze the proteomes of organisms. MudPIT is a chromatography-based proteomic technique where a complex peptide mixture is prepared from a protein sample and loaded directly onto a triphasic microcapillary column packed with reversed phase, strong cation exchange, and reversed phase HPLC grade materials. Once the complex peptide mixture is loaded onto the triphasic microcapillary column, this column is placed directly in-line with a tandem mass spectrometer. The tandem mass spectrometry data generated from a MudPIT run is then analyzed to determine the protein content of the original sample. MudPIT has been proven to be an excellent tool for both qualitative and quantitative proteomic analyses.

Quantitative Proteomics Method Development

The dynamic changes of a proteome or fractions of a proteome; i.e., organelles and protein complexes, can be analyzed via quantitative proteomic methods. We largely carry out label free quantitative proteomic analyses using spectral counting. In spectral counting, the total number of tandem mass spectra that match peptides to a particular protein is used to measure the abundance of proteins in a complex mixture. We have developed the normalized spectral abundance factor (NSAF) approach for using spectral counting in quantitative proteomics (Zybailov et al., 2006). This approach takes into account the sample-to-sample variation that is obtained when carrying out replicate analyses of a sample and the fact that longer proteins tend to have more peptide identifications than shorter proteins. Examples of the application of the NSAF approach to quantitative proteomic analysis include work on the expression changes of membrane proteins in S. cerevisiae (Zybailov et al., 2006) and on the human transcriptional regulatory complex, Mediator (Paoletti et al., 2006). We have also demonstrated that the NSAF approach generates datasets that have a high degree of statistical similarity to Affymetrix GeneChip™ datasets (Pavelka et al., 2008). This opens the door for NSAF based MudPIT analyses to carry our similar studies to those done over the years for GeneChip™ analyses and provides a foundation for bioinformatics analysis of such datasets using established GeneChip™ tools (Pavelka et al., 2008).  We have recently improved the NSAF approach and we now use the distributed NSAF (dNSAF) approach for our quantitative proteomics analyses (Zhang et al., 2010).  In the dNSAF approach, peptides/spectral counts that are shared between proteins are distributed to each protein based on the amount of unique peptides/spectral counts found for each protein. 

Dynamics of Multiprotein Complexes         

Our research is focused on multiprotein complexes. Using affinity purification coupled with MudPIT, the NSAF, and now the dNSAF approach, we analyze complexes using different tagged subunits from the same multiprotein complex. This allows us to determine the relative abundance of particular proteins in a complex in a bait-dependent fashion and leads to the analysis of distinct forms of multiprotein complexes that can have important functional insights. In addition, current projects in the lab are using these approaches to determine the impact of different stimuli on multiprotein complexes.

Quantitative Proteomic Analysis of Transcriptional Regulatory Complexes

After affinity purification to purify protein complexes from cells, we use the MudPIT and the NSAF, and now dNSAF approach to quantitatively analyze multiprotein transcriptional regulatory complexes. The combination of these approaches provides not only a list of the proteins present but also the abundance of proteins present. This information can be used in a variety of ways. First, a quantitative proteomic analysis of a protein complex can be carried out to gain insight into the forms of the protein complexes and whether or not there are any abundant and poorly characterized new protein-protein interactions.

Examples of our use of this approach include a quantitative proteomic analysis of HeLa cell Mediator where four biological replicates of four different Mediator subunits were used to affinity purify the complex (Paoletti et al., 2006). The technology described in this published work is an important foundation for our research where biological replicates of transcriptional regulatory complexes and subassemblies are analyzed in a bait dependent manner and the statistically significant differences in the complexes and subassemblies are determined.

We have begun to apply this approach to RNA polymerase I, II, and III. During the course of this work, we determined the function of a poorly characterized but highly conserved protein, Rtr1, which was interacting with RNA polymerase II (Mosley, Pattenden, et al., 2009). We then pursued the function of Rtr1 in the transcription cycle and found that Rtr1 localized in coding regions between the peaks of Serine5-phosphorylation and Serine2-phosphorylation of the C-terminal domain of the largest subunit of RNA polymerase II, Rpb1.  Deletion of Rtr1 resulted in an accumulation of the Serine5-phosphorylated form of Rpb1, a decrease in RNA polymerase II transcription, and termination defects. We further demonstrated that Rtr1 is a Serine 5 phosphatase of Rpb1 C-terminal domain (Mosley, Pattenden, et al., 2009). We have also analyzed yeast RNA polymerase I, II, and III using the quantitative proteomics approaches developed in the lab.  We demonstrated that we can properly cluster the proteins in these complexes, determine the relative abundance of proteins in the complexes, and we are able to capture dynamic interactors of the complexes (Mosley et al., 2011)

Probabilistic Protein Interaction Network Analysis

We have a growing interest in protein interaction network analyses. As a result of ongoing collaborations with other principal investigators at the Stowers Institute, we analyze a large amount of diverse affinity purifications from organisms like S. cerevisiae and human tissue culture. We used a portion of this data to develop a novel approach for assembling probabilistic local protein interaction networks using vector algebra and statistical methods, and applied this to the human Tip49a/Tip49b protein interaction network (Sardiu et al., 2008). We defined four protein complexes, URI/Prefoldin, hINO80, SRCAP, and TRRAP/TIP60 and we identified new components of these complexes. Most importantly, we determined the probabilities of protein-protein interactions within and between complexes. We used NSAF values to determine the probability of each protein-protein interaction in the dataset. In a limited follow up analysis, higher probabilities corresponded to positive coIPs and low probabilities corresponded to negative coIPs (Sardiu et al., 2008). These results raise the intriguing possibility that the probabilities that are calculated from this technology may provide insight into the architectural organization of a protein complex. We have further analyzed the Tip49a/Tip49b dataset to evaluate different clustering algorithms to gain insight into the potential value of different clustering approaches for future computational assembly of unknown protein complexes (Sardiu et al., 2009). We are currently developing a large dataset protein-protein interactions with a large number of affinity purifications of proteins involved in transcription. Future studies will be to assemble and analyze these datasets into probabilistic protein interaction networks.

Modularity of Multiprotein Complexes and Protein Complex Architectures

We have begun to utilize deletion network analyses to propose the architecture of multiprotein complexes and to determine the modularity of protein complexes.  In our initial publication in this area, we analyzed the Rpd3/Sin3 histone deacetylase complexes.  Here, Rpd3 was used as an affinity purification bait and protein complexes were analyzed from a wild type background and in strains where different subunits of the complex where deleted.  Using this approach we were able to propose an architectural model for the Rpd3/Sin3 histone deacetylase complexes (Sardiu et al., 2009).  More recently, we have expanded this approach for the analysis of the SAGA and ADA chromatin remodeling complexes (Lee et al., 2011).